Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A system for evaluating a classifier implemented within an image signal processor (ISP) comprising: a microprocessor and memory storing a training image set having a plurality of images, the microprocessor being capable of sending each of the plurality of images to the ISP, the plurality of images including Mtotal images indexable as 1, 2, . . . M; and machine-readable instructions stored within the memory and executed by the microprocessor capable of: selecting, based upon a divider position within the training image set denoted by a divider integer D between one and M+1 inclusive, the D th image of the plurality of images; controlling the ISP to classify the D th image as belonging to or not belonging to an object class; determining a positive-match count equal to zero when the D th image is classified as not belonging to the object class and equal to one when the D th image is classified as belonging to the object class; determining an error count based upon (a) total number of images of the training image set belonging to the object class, and (b) the positive-match count; repeating, for a plurality of other divider positions within the training image set, the steps of selecting, controlling, and determining to identify an optimal divider position corresponding to at least one of (a) a minimum-error count and (b) a maximum-error count, wherein total number classifier operations performed by the classifier to determine at least one of the minimum-error count and the maximum-error count equals total number of images in the training image set; and determining optimality of the classifier by comparing at least one of (a) optimal divider position corresponding to a minimum-error count to a predetermined optimal divider position corresponding to a predetermined minimum-error count, and (b) optimal divider position corresponding to a maximum-error count to a predetermined optimal divider position and corresponding to a predetermined maximum-error count.
A system for evaluating an image classifier within an image signal processor (ISP) uses a microprocessor and memory. The memory stores a set of training images (M total). The microprocessor sends each image to the ISP. Software instructions, when executed, select an image within the training set based on a divider position (D). The ISP then classifies the image as belonging or not belonging to a specified object class. A positive-match count is determined: 1 if the image belongs to the object class, 0 otherwise. An error count is calculated based on the total number of images belonging to the object class and the positive-match count. This process repeats for other divider positions to find an optimal divider position (minimum or maximum error). The total number of classifier operations equals the number of training images. Finally, the classifier's performance is evaluated by comparing the optimal divider position to a known optimal divider position and corresponding minimum/maximum error count.
2. The system of claim 1 , the machine-readable instructions being further capable of: determining a negative-match count equal to one when the D th image is classified as not belonging to the object class; and determining the error count as proportional to the positive-match count subtracted from sum of (a) the negative-match count and (b) the total number of images of the training image set belonging to the object class.
The image classifier evaluation system from the previous description also calculates a negative-match count: 1 if the Dth image is classified as *not* belonging to the object class. The error count is then calculated as proportional to the positive-match count subtracted from the sum of the negative-match count and the total number of training images belonging to the object class. In other words, the error calculation now explicitly considers both positive and negative matches for a more refined error rate.
3. The system of claim 1 , the machine-readable instructions being further capable of: determining, for the D th image where D>1, the error count as a cumulative error count partially based upon one previously-determined positive-match count and one previously-determined negative-match count for the (D−1) st image of the plurality of images.
The image classifier evaluation system from the first description determines the error count as a cumulative error count partially based upon a previously-determined positive-match count and a previously-determined negative-match count for the (D-1)th image of the plurality of images, specifically when D is greater than 1. Instead of independently computing the error for each image, the system leverages the error calculated for the previous image to improve efficiency or to implement a specific error accumulation model.
4. The system of claim 1 , wherein total number classifier operations performed by the classifier to determine at least one of the minimum-error count and the maximum-error count is a linear function of total number of images in the training image set.
In the image classifier evaluation system from the first description, the total number of classifier operations performed to determine the minimum or maximum error count scales linearly with the total number of training images. This means the processing time increases proportionally to the size of the training dataset, indicating a predictable and manageable performance characteristic of the evaluation system.
5. The system of claim 1 , the plurality of other divider positions being each other possible divider position.
In the image classifier evaluation system from the first description, the system repeats the evaluation process for *every* possible divider position within the training image set. This exhaustive search ensures that the determined optimal divider position is truly the best among all possibilities, rather than being limited by a smaller, potentially biased, subset of divider positions.
6. A method for evaluating a classifier implemented within an image signal processor (ISP) to identify an object class in a received electronic image signal, the method comprising: selecting, from a training image set having a plurality of images including M total images indexable as 1, 2, . . . , M, the D th image of the plurality of images where D denotes a divider position within the training image set; controlling the ISP to classify the D th image as belonging to or not belonging to an object class; determining a positive-match count equal to zero when the D th image is classified as not belonging to the object class and equal to one when the D th image is classified as belonging to the object class; determining an error count based upon (a) total number of images of the training image set belonging to the object class, and (b) the positive-match count; repeating, for a plurality of other divider positions within the training image set, the steps of selecting, controlling, and determining to identify an optimal divider position corresponding to at least one of (a) a minimum error count and (b) a maximum-error-count, wherein total number classifier operations performed by the classifier to determine at least one of the minimum-error count and the maximum-error count equals total number of images in the training image set; and determining optimality of the classifier by comparing at least one of (a) optimal divider position corresponding to a minimum-error count to a predetermined optimal divider position corresponding to a predetermined minimum-error count, and (b) optimal divider position corresponding to a maximum-error count to a predetermined optimal divider position and corresponding to a predetermined maximum-error count.
A method for evaluating an image classifier within an image signal processor (ISP) identifies an object class in an electronic image. The method selects an image from a training set (M total), based on a divider position (D). The ISP classifies the image as belonging/not belonging to an object class. A positive-match count is determined: 1 if it belongs, 0 otherwise. An error count is calculated based on the total object-class images and the positive-match count. These steps repeat for other divider positions to identify the optimal divider position (minimum or maximum error). The total classifier operations equal the total number of images. Finally, the classifier's performance is determined by comparing the optimal divider position to a known optimal divider position and corresponding minimum/maximum error count.
7. The method of claim 6 , further comprising: determining a negative-match count equal to one when the D th image is classified as not belonging to the object class; and determining the error count as proportional to the positive-match count subtracted from sum of (a) the negative-match count and (b) the total number of images of the training image set belonging to the object class.
The image classifier evaluation method from the previous description also calculates a negative-match count: 1 if the Dth image is classified as *not* belonging to the object class. The error count is then calculated as proportional to the positive-match count subtracted from the sum of the negative-match count and the total number of training images belonging to the object class. In other words, the error calculation now explicitly considers both positive and negative matches for a more refined error rate.
8. The method of claim 6 , the step of determining the error count further comprising, for the D th image where D>1, determining the error count as a cumulative error count partially based upon one previously-determined positive-match count and one previously-determined negative-match count for the (D−1) st image of the plurality of images.
The image classifier evaluation method from the sixth description determines the error count as a cumulative error count partially based upon a previously-determined positive-match count and a previously-determined negative-match count for the (D-1)th image of the plurality of images, specifically when D is greater than 1. Instead of independently computing the error for each image, the method leverages the error calculated for the previous image to improve efficiency or to implement a specific error accumulation model.
9. The method of claim 6 , the step of selecting further comprising selecting the D th image such that total number classifier operations performed by the classifier to determine at least one of the minimum-error count and the maximum-error count is a linear function of total number of images in the training image set.
In the image classifier evaluation method from the sixth description, the Dth image is selected such that the total number of classifier operations performed to determine the minimum/maximum error count scales linearly with the number of training images. This ensures that the processing time of the evaluation method increases predictably with the size of the training dataset.
10. The method of claim 6 , in the step of repeating, the plurality of other divider positions being each other possible divider position.
In the image classifier evaluation method from the sixth description, the method repeats the evaluation process for *every* possible divider position. This exhaustive search ensures that the determined optimal divider position is truly the best among all possibilities, rather than being limited by a smaller, potentially biased, subset of divider positions.
Unknown
December 12, 2017
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.